EP1271956A1 - Method and system for the conversion of XHTML/HTML documents to MPEG-4 documents - Google Patents

Abstract

The invention is related to a device to convert HTML data into MPEG-4 data comprising

• means to generate a structure of the HTML data and of the HTML data linked to said HTML data,

According to the invention, the device comprises:

• means to convert the HTML page and frame layout into MPEG4 layout,
• means to convert the HTML data into MPEG-4 nodes.

Description

• The invention concerns a device and a method to convert HTML (standing for "Hypertext Markup Language") data into MPEG-4 (MPEG standing for "Motion Picture Expert Group") data.
• More generally, the domain of the invention is the domain of multimedia.
• For many years, the use of Internet for the distribution of information has been done using data in a standard format called HTML and now XHTML (standing for "Extensible Hypertext Markup Language"). This format being a textual format, its transmission is not efficient.
• Throughout the whole text, HTML stands also for XHTML.
• Meanwhile, the MPEG working group has been working on the compression and transmission of audio-visual content in MPEG-1 and MPEG-2 standards. The first one was mainly dedicated to the containment of video films on a single CD-ROM called video-CD. The quality was poor but acceptable for the computing capacity of personal computers. MPEG-2 is mainly dedicated for the broadcast of audio-visual data on wideband platforms (cable, satellite, etc.).
• The global bandwidth capacity of Internet being improved since a few years has made possible the convergence of digital TV and Internet. The standard enabling such transmission is called MPEG-4.
• MPEG-4 standard does not specify the transport means of MPEG-4 information but specify how data have to be compressed and encapsulated in order to be composed on an MPEG-4 platform.
• MPEG-4 data are of different types and can be:
• binary compressed data, called BIFS, (standing for Binary Format for Scene) sent through streams.
• audio data, video data in a natural way but also in a synthetic form : 2D and 3D geometrical objects, synthetic sound, etc..
• These data are composed into a scene. MPEG-4 enables the description of user interactivity and objects animation and updates.
• The MPEG-4 objects are composed of nodes describing the MPEG-4 scene. Nodes can represent geometrical objects, pictures, sounds or videos.
• XHTML/HTML documents are standardized textual files. They are composed of elements described into HTML DTD (standing for "Document Type Definition") and CSS (standing for "Cascading Style Sheets") that can be found in the HTML specification document.
• All these HTML elements define the structure of the XHTML/HTML page. Such a page can contain or can include embedded links to other web pages, pictures, sounds and videos. There is also a limited interactivity language called JavaScript that can be embedded in the document.
• The development of Internet and of the multimedia has increased the amount of data exchanged between terminals of any kind such as personal computers, set-top boxes. Even if the general Internet bandwidth has been increased, the volume of data transmitted increases also in a considerable manner as more and more users are connected to the Internet and more and more complex and high quality data are exchanged.
• Moreover, the development of multimedia applications is linked to the development of Internet, the World Wide Web being a common source of multimedia data.
• So, the invention increases the convergence of multimedia and of Internet by providing a device enabling to display the content of HTML information on an MPEG-4 platform.
• The object of the invention is a device to treat HTML information including data and at least one page and its associated frame layout, comprising
• means to generate a structure of the HTML data,
• According to the invention, the device comprises
• means to convert the HTML page and its associated frame layout into MPEG4 layout,
• means to convert the HTML data into MPEG-4 nodes.
• The HTML formats and MPEG-4 formats do not address the same types of terminals and thereby are completely different, especially in terms of layout. By converting the HTML data into MPEG-4 data, the device enables to solve the problems of displaying HTML data on MPEG-4 terminals and especially in term of layout and visual coherence.
• This device enables the user to display HTML data without using a specific browser and then the merging of native HTML data and of MPEG-4 data can be done
• Having a single browser for both MPEG-4 and XHTML/HTML data can allow not disturbing the user by switching from a content to another by using two different interfaces.
• The means to generate a tree structure of the HTML data are able to generate a tree structure of the associated linked data.
• Moreover, on the World Wide Web, there are much more HTML data than MPEG-4 data, so that it is particularly usefull to provide a device enabling the conversion of HTML data in MPEG-4 data on platforms dedicated to display data issued from the World Wide Web.
• Preferentially the means to generate a structure of the HTML data generate a tree structure of the HTML data are able to generate a tree structure of the HTML data and that the conversion means intended to convert the HTML data into MPEG-4 nodes convert the HTML tags and the cascading style sheets to Binary Information For Scene information.
• The elements and the structure of these elements are different in MPEG-4 and in HTML so the device must have the means to convert the elements of the HTML document in MPEG-4 nodes and then it must also have the means to convert the page and frame layouts of the HTML document as the MPEG-4 format does not have neither frame nor page concept. This module converts the frame and page layouts of the HTML document in the MPEG two-dimensional layout so that it is possible to keep the consistency between the different media (audio, video,..) in the MPEG-4 scene and to keep the visual aspect.
• Advantageously, the HTML data are standardized HTML data or browser-specific HTML data.
• In this manner, any type of data can be considered and converted using the device according to the invention. Indeed, usually, the HTML files are written in view of being viewed using one dedicated type of browser or at least are preferentially viewed using certain proprietary browsers, so written using dedicated proprietary elements.
• In one embodiment, the device comprises means to generate an output document from the obtained MPEG-4 layout and MPEG-4 nodes, the output document being described using the MP4 format or the XMT format.
• This gives the possibility to generate a file for different purposes and not only for displaying. Namely the XMT format is a textual file and MP4 is a binary format. Some elements of the HTML document can be converted in a different way according to the chosen output file. For instance, the comments included in the HTML file can be ignored by the converter if the output file is an MP4 file and can be converted into comments if the output file is an XMT file.
• The invention is also related to a multimedia system. According to the invention, it comprises :
• at least one browser to generate a request of at least one HTML document located on a remote server,
• at least one device according to any of the embodiments of the invention,
• at least one transmission link between said remote server and said device according to any of the embodiments of the invention for transferring said requested HTML document,
• at least one display to display said MPEG-4 document, said device according to any of the embodiments of the invention to convert said requested HTML document into an MPEG-4 document before transferring it to said display.
• The invention is also related to a multimedia system characterized in that it includes
• a server,
• a device for displaying multimedia data,
• a device for receiving a request for an HTML document from a user,
• a transmission link being able to for connecting said server and said device for receiving the request for transmitting said request to said server,
  said server including a device according to any of the embodiments of the invention for converting said HTML document into an MPEG-4 document, said transmission link transmitting said HTML document converted into MPEG-4 to the device for displaying multimedia data.
• This embodiment is particularly interesting. The inventors have noticed that as MPEG-4 data are heavily compressed, it is better to transmit textual information compressed into MPEG-4 format than to transmit HTML data. Indeed it saves bandwidth on the transmission link.
• In a preferred embodiment, the server is connected to the World Wide Web and said requested HTML document is located on the World Wide Web.
• The invention is also related to a method for treating HTML information including at least one page and its associated frame layout wherein
• one generates a structure of the HTML data and of the HTML data linked to said HTML data.
• According to the invention,
• one converts the HTML page and frame layout into MPEG4 layout,
• one converts the HTML data into MPEG-4 nodes.
• The present invention also relates to a computer program product comprising portions of software code or program instructions adapted to implement the method of claim 8 for converting HTML data into MPEG-4 data, when the computer program is executed on a computer.
• A computer program product can be any computer program stored in a memory of any kind such as a CD-Rom, a floppy disk, or also a computer program downloaded from a network.
• Other characteristics and advantages of the invention will appear through the description of a non-limiting embodiment of the invention, which will be illustrated, with the help of the enclosed drawings among which
• figure 1 represents a first embodiment of a multimedia system according to the invention,
• figure 2 represents a second embodiment of a multimedia system according to the invention,
• figure 3 represents the splitting up of an HTML document into an MPEG-4 document
• On figure 1, a server 7 includes a device 1 for converting HTML data into MPEG-4 data. The server 1 is connected to the World Wide Web 5 through transmission link 6.
• Said server is also connected through transmission link 8 to the device 2 including a browser 3. The device 2 is also connected to a target platform 4. In a preferred embodiment the target platform is a display.
• The device 2 can be, in a preferred embodiment, a set-top-box. It is called in this document a user terminal. The user terminal 2 includes a user interface 3, also called browser. This browser enables the user to select a multimedia file on a remote server; this multimedia file is preferentially located on the World Wide Web. The browser allows the user to select different coded types of files. The user can select native MPEG-4 data to display them on its display 4, the user can also select other coded types of files such as HTML data, thereby using the same interface.
• The user sends a request for one data to the browser 3. Then, the request is transmitted to server 7 through transmission link 8. This transmission link can be for example an Ethernet link.
• In the embodiment represented on figure 1 the requested HTML document is located on a server connected to the World Wide Web.
• When the requested document is an HTML document, the server 7 receives the request. Depending on the location of the requested file (specified by its URL (standing for "uniform Resource Locator") address) the request is transmitted to the appropriate server on the World Wide Web through transmission link 6. The requested document and if needed, the documents linked to this requested document, are received on the server 7 and transmitted to the device 1. Device 1 converts the HTML documents into an MPEG-4 document according to the invention and detailed later in this document. The converted document is then transmitted to the user terminal 2 and transmitted to the display 4.
• In this embodiment, one advantage is to have a centralized server, which makes the format conversion from HTML to MPEG-4. This avoids implementing a device according to the invention in any user terminal.
• A second advantage is to transmit MPEG-4 data between the server 7 and the user terminal 2. Indeed, MPEG-4 data are compressed data whereas HTML data are not compressed, so the data rate is improved and bandwidth is saved.
• Figure 2 represents another embodiment of a multimedia system according to the invention.
• A user terminal 2 is connected to the World Wide Web 5 through a transmission link 6.
• In this embodiment, a device 1 for converting HTML data into MPEG-4 data is located in the user terminal 2. There is no global server for centralising the conversion of HTML data into MPEG-4 data.
• When a user requests an HTML document, the user terminal 2, which is connected to the World Wide Web, sends the request to the World Wide Web. The document, while received on the user terminal 2, is converted into MPEG-4 format in device 1 as described later in this document. Then the converted document is transmitted to the target platform display 4.
• This embodiment has the advantage to provide to each user a device according to the invention even in a heterogeneous environment, it means no global server is needed.
• In a particular embodiment, the device 2 is a computer including a processor and memory, said memory comprising portions of software code or program instructions adapted to implement a method for converting HTML data into MPEG-4 data as described on figure 3, when the computer program is loaded into the processor. Said memory can be non-volatile memory or volatile memory or also a hard disk, a floppy disk or a CD-Rom.
• The format conversion of an HTML document 9 is represented on figure 3.
• An HTML document can be composed of text and Cascading Style Sheet 10 (referenced as CSS on figure 3), images, audio and video objects 12, script 13, forms 14, or other data format such as DHTML 15 (standing for "Dynamic Hypertext Markup Language"), Java data or VRML (standing for "Virtual Reality Modelling Language").
• Cascading Style Sheets 10 describe how documents are presented on screens, in print, or perhaps how they are pronounced. By attaching style sheets to structured documents on the Web (e.g. HTML), authors and readers can influence the presentation of documents without sacrificing device-independence or adding new HTML tags. As XHTML extends HTML with XML language, XSL extends CSS in the same way.
• The two languages complement each other and can be used together.
• The text and Cascading Style Sheets 10 are converted using a BIFS conversion module 16. The images, audio and video objects can be converted into MPEG-4 using an MPEG-4 format transcoder module 17. The other data related to DHTML, Java, VRML 15 can be converted using another conversion module 18 into BIFS or MPEG-J. Conversion module 16, transcoder module 17 and MPEG-J conversion module 18 are linked in the way they have to share informations for Object Descriptor settings 20.
• The device 1 converts the text, cascading style sheets into MPEG-4 nodes keeping the same visual aspect, namely converting also the layout.
• When the device 1 receives the HTML requested document it parses all the pages links. In case the linked media or document is needed for further treatment, like display information, it downlads it or keeps reference on it. Specifically, for transcoding of images or sounds for example, media type and specifications have to be kept for Object Descriptor initialization.
• The common structure of an HTML document can be the following:

  
• All the tags have opening tags and closing tags. An opening tag is for example <HTML> and the corresponding closing tag is </HTML>. However, in some cases, the closing tags are omitted. In this case, the converter can however make the conversion and does not generate any error. The converter makes the conversion by supposing the location of the missing closing tags.
• We are now going to describe some specific elements of an HTML page.
• The "!DOCTYPE" element identifies the XHTML/HTML document type. The converter checks if it manages the document type. If the document type is not correct or not manageable, the converter can suppose that either the document is an HTML document and then it tries to begin the conversion or that the document is not in HTML and then it can stop the current conversion and display an error message.
• The converter initializes its internal MPEG-4 structure to a top Layer2D node, which defines the initial window frame, with default sizes (-2 for width or X size, -2 for height or Y size).
• In XHTML/HTML, colors are described in two different ways:
• The first way is the hexadecimal description: for example "#FF9F00". This needs a specific conversion for MPEG-4 color:
• R (Red component): ((color & 0xFF0000) >> 16) / 255.
• G (Green component): ((color & 0xFF00) >> 8) / 255.
• B (Blue component): (color & 0xFF) / 255.
• This conversion corresponds to the application of a bit mask to the HTML color ASCII value to extract red, green and blue components and a division to range them within the MPEG-4 color values.
• The second way is the textual description: for example "black". These descriptions correspond to fixed hexadecimal colors. For example, "black" corresponds to "#000000" in HTML and to "0 0 0" in MPEG-4.
• The "!--" element corresponds to an HTML comment. This comment is converted into an MPEG-4 comment or ignored depending on the chosen output file. If the output file is an XMT output file, this comment can be converted into an MPEG-4 comment, if the output file is an MP4 output file, this comment is ignored.
• The magic targets define in MPEG-4 the frames into which the linked HTML content has to be displayed.
• "_blank": The converter creates a new Layer2D (with default size) on top of all other contents. One can also replace the complete scene by the new Layer2D.
• "_self": The converter replaces the current Layer2D content by the new one.
• "_parent": The converter replaces the parent Layer2D content by the new one.
• "_top": The converter has to replace the top Layer2D content by the new one.
• These Layer2D can also be called here visual frames to make an analogy with HTML principles.
• The use of such target in hyperlinks needs specific treatment like BIFS Update command, in a Conditional node for example. Another solution to manage such behavior is to put an Inline node as children of all Layer2D (frame) and then to update the content of this Inline dynamically.
• A specific use of these targets is to define specific place in the HTML document thanks to the '#' character and to the HTML 'name' field. This implies having an additional Transform2D as top children of all Layer2D in order to be able to position the converted page as requested. Additionally, many Valuator or Conditional nodes are necessary to update these Transform2D translation, and PlaneSensor2D offset, in case of use of this node as scrolling feature.
• XHTML/HTML defines 7 different font sizes. These font sizes are only references and do not apply to explicit display sizes. Most web browsers use a system font size of 12 as the default HTML font size (which is 3). The display sizes equivalent to these HTML font sizes, are chosen considering the display type and the capabilities of the target platforms.
• The initial application settings (BIFS decoderConfigDescriptor pixelWidth, pixelHeight, pixelMetric, ...) in the IOD (standing for "Initial Object Descriptor") sizes should be used for the converter sizes but these can also be manually set.
• The converter knows the display resolution (width and height) of the target platform 4 in order to be able to position the objects either dynamically or statically in the MPEG-4 scene.
• In XHTML/HTML, sizes can be set using several different ways:
• pixel aligned. Example: "width=50".
• relative size. Example: "height=15%".
• Multiple, auto-size. Examples: "width=1*", "height=*".
• As these sizes are not compliant with MPEG-4 size ranges and definition, they need to be converted to MPEG-4 content size, depending on the target platform application settings (display size and capabilities). This can need pixel to relative size conversion. In most cases, the converter stores global sizes, especially in the third size case.
• In case the converted frames (Layer2D) width (x size) and/or height (y size) cumulative content sizes are bigger than the target platform application display sizes, the converter introduces x and/or y scrolling capabilities in the layers thanks to layer dragging or insertion of scrolling bars. This enables the user to view the whole document.
• Concerning the conversion of text, the following table gives a list of HTML elements and their corresponding converted MPEG-4 nodes. This table describes conversion of single HTML elements but some can need multiple and linked MPEG-4 nodes to restore the same functionality. Moreover, some HTML elements have cumulative effects, such as font properties, and then the converter keeps status of these effects.
• Some linked documents or media can be described in other format than in HTML. Specifically, when JavaScript is defined as function of field, like 'onblur', this is ignored by the converter unless it can manage it by a specific way (by MPEG-J conversion for example). In case the converter does not manage such conversion, a message is displayed at the location where the non convertible element should be placed.
• The table in the annex 1 represents the conversion of some HTML elements into MPEG-4 nodes. This table gives also the method used for the conversion. The list given in this annex is non limitative.
• When the verbs 'contain' or 'embed' are used in this table it means that the specified parent node possesses the given node in its children field.
• In all the tables of the different annexes, MPEG-4 nodes are given in italic bold and MPEG-4 field node in italic. For example, Layer2D node and its size field.
• An HTML document contains also some proprietary elements that have to be converted. These proprietary elements can be for instance specific to the browser used. They are not standardized and have different names according to the browser used.
• The table given in annex 2 illustrates the conversion used for some of these proprietary and non standard elements. This table is not an exhaustive list of the proprietary elements.
• The Cascading Style Sheet langage reference introduces the possibility to setup the default display parameters such as the color, the font. These parameters are stored and used by the converter for display purpose. The conversion of Cascading Style Sheet parameters is indicated in the annex 3. The annex 3 gives a non exhaustive list of the Cascading Style Sheet parameters.
• The output file 19 can be delivered to an MPEG-4 system (client or server). This can be an MP4 binary file or can be an XMT textual file. The output media is set according to the target platform 4. If the target platform is a display, the output file 19 is an MP4 file, if the target platform is a different device, the output file 19 can be an XMT file.
Annex 1

• XHTML/HTML elements and parameters (attributes)	MPEG-4 node - method to convert
  A		The media (Text, image, etc.) contained in this element are children of an Anchor node.
  	charset	Character encoding used for the address. NB: Only UTF-8 is understood by MPEG-4.
  	type	Useful to know if linked content type can be understood by the converter before it downloads it.
  	name	This feature needs specific treatment in order to manage it: the use of a Transform2D node on top of the embedding Layer2D is necessary as well as a Valuator node in order to position the page as requested.
  	hreflang	Useful to know if linked content language can be understood by the converter.
  	rel	Useful if the converter module manages pre-load of contents.
  	rev	Useful if the converter module manages pre-load of contents.
  	accesskey	The Anchor node can be triggered by a KeySensor. This node has not yet been standardized. This functionality is subject to change.
  	shape	This is the transparent children Shape of this Anchor.
  	coords	These are the properties of the geometry of the previous Shape.
  	tabindex	Such functionality need specific conversion (MPEG-J or other)
  	target	This is the param of the Anchor node. As discussed in the document, when the 'name' or 'target' features are used, this implies all Frame2D to contain an Inline node which DEF name is given by this parameter. This allows to change the URL when the content is updated by an hyperlink.
  	onfocus	This Anchor contains a TouchSensor with ROUTE to its isOver.
  	onblur	This Anchor contains a TouchSensor with ROUTE to its isOver.
  	href	Address (url and param fields) of the Anchor node.
  ABBR		This is a Text node, or the continuity of the string field of the previous one. This should also be put in the WorldInfo node.
  ACRONYM		This is a Text node, or the continuity of the string field of the previous one. This should also be put in the WorldInfo node.
  ADDRESS		This is a Text node, or the continuity of the string field of the previous one, with an ITALIC FontStyle display. This information should also be included in the WorldInfo node.
  APPLET		This shall be ignored and the HTML alternate message should be displayed instead
  AREA		This is an Anchor node. Any Shape that it contains have to be transparent.
  	shape	This is the transparent Shape contained in this Anchor.
  	coords	These are the properties of the geometry of the previous Shape.
  	href	Address (url and param) of the Anchor node.
  	nohref	This Anchor does not link (no url)
  	alt	This is the description field of the Anchor node.
  B		This is a new Text node, in the continuity of the previous one, if it exists, with a BOLD FontStyle display and cumulative with the style of the previous Text node.
  BASE		This is not a visual information. This has to be used by the converter for a different purpose.
  	href	This means that all URLs in the XHTM/HTML document are based on the given URL. This is not a visual information but address information for the converter.
  	target	This means that all hyperlinks (Anchor nodes) have to be displayed in a specific Layer2D. When the 'name' or 'target' features are used, this implies that all Frame2D must contain an Inline node which DEF name is given by this parameter. This allows to change the URL when the content is updated by an hyperlink.
  BASEFONT		These parameters have to be kept by the converter as default Material2D, FontStyle for Text nodes. This element is cumulative (i.e. it can be called several times in the web page to update values).
  	id	This is the language of the text, can be useful at application level, for language switching.
  	size	This is the default size field of the FontStyle. This size is in HTML range (1-7) and then have to be converted in pixel or relative size.
  	color	This is the default emissiveColor value of the Material2D of the Shape embedding the Text node. The HTML color can be set with an hexadecimal value, then should need conversion.
  	face	This is the default family field value of the FontStyle.
  BDO		This implies changes on the LeftToRight field of the FontStyle node.
  	lang	This is the language of the text, can be useful at Application level, for language switching.
  	dir	If equals to "LTR" then LeftToRight is TRUE, FALSE otherwise.
  BIG		This increases the size field of the FontStyle Node. The increase is of 1 in HTML, it has then to be recomputed in pixel or relative size. A new Text node has then to be created in the continuity of the previous one if it exists thanks to the Form node.
  BLOCKQUOTE		This creates a new Text node with same FontStyle than the previous one if it exists. This text is indented and has a blank like before and after its sentences.
  BODY		This delimitates the HTML header from the body (visual part). A Layer2D node has to be created.
  	alink	This is the default emissiveColor value of the Material2D of the Shape embedding the Text node of an Anchor node on which the TouchSensor isActive field is TRUE. The HTML color can be set with an hexadecimal value, then should need conversion.
  	background	This is the url field of the Background2D node contained in the created Layer2D.
  	bgcolor	This is the backColor field of the Background2D node contained in the created Layer2D.
  	bgproperties / nohtilebg / novtilebg	If "FIXED", the image of the Background2D is not linked by a ROUTE to the layer movements (horizontally, vertically or both).
  	credits	This can be used in the WorldInfo node.
  	fontsize	This is the default size field for all FontStyle nodes embedded in the Layer2D. This size is in HTML range (1-7) and then has to be converted in pixel or relative size.
  	hspace	This is a value for alignment in pixel for horizontal constraints with the borders in the Form node contained at the top of the Layer2D node and for all nodes embedded in it.
  	instructions	Can be used as the Anchor for a help button in the scene.
  	leftmargin / marginwidth	This has the same effect than the hspace field.
  	link	This is the default emissiveColor value of the Material2D of the Shape embedding the Text node of an Anchor node. The HTML color can be set with an hexadecimal value, then should need conversion.
  	logo	This can be used at application level as application icon.
  	vspace/ marginheight/ topmargin	This has the same effect than the hspace field but for vertical borders.
  	onfocus	This Layer2D contains a TouchSensor with ROUTE to its isOver.
  	onblur	This Layer2D contains a TouchSensor with ROUTE to its isOver.
  	onload	This Layer2D contains a VisibilitySensor with ROUTE to its enterTime.
  	onunload	This Layer2D contains a VisibilitySensor with ROUTE to its exitTime.
  	text	This is the default emissiveColor value of the Material2D of all Shape nodes embedding any Text node. The HTML color can be set with an hexadecimal value, then should need conversion.
  	vlink	This is the default emissiveColor value of the Material2D of the Shape embedding the Text node of an already selected Anchor node. The HTML color can be set with an hexadecimal value, then should need conversion. This can only be implemented if the converter can store a certain number of already visited links to get their status.
  	xspeed/ yspeed	This is the speed in pixel per second for the movement of the background image. This functionality needs a Layout node over the Background2D of the current Layer2D.
  BR		g field of the current Text node or the creation of a lent one.
  BUTTON		This creates a graphical button which can be represented by a picture in MPEG-4. The size of the button will depend of the image or text it is displayed on it. Then there is a need of a Form node over the Shape for this ImageTexture. A TouchSensor is embedded in the same Shape as this button.
  	accesskey	A KeySensor is embedded in the same Shape than this button. This functionality is not yet defined in MPEG-4 and then is subject to change.
  	disabled	The TouchSensor enabled field is set to FALSE.
  	name	The Shape node is named. Use a DEF.
  	tabindex	This cannot be managed by MPEG-4 BIFS. This can be managed by MPEG-J for Keyboard control.
  	type	Only take into account the "button" type otherwise the converter has to ignore the button.
  	onfocus	The TouchSensor has a ROUTE to its isOver.
  	onblur	The TouchSensor has a ROUTE to its isOver.
  	value	The converter has to ignore this field.
  CAPTION		This is a table caption, only valid in a table HTML element. This is a Text node which FontStyle style field is set to "ITALIC".
  	align	This functionality needs a Form node over this Text node and the MPEG-4 conversion of the table. This is an horizontal constraint between the table and the Text node.
  	valign	This functionality needs a Form node over this Text node and the MPEG-4 conversion of the table. This is a vertical constraint between the table and the Text node.
  CENTER		This implies a Form node with horizontal centering constraint for following converted HTML elements. This can also be managed statically by the converter that knows display settings (size) and then can position the objects by simple X and Y computation.
  CITE		This is a new Text node, in the continuity of the previous one if it exists, with an ITALIC FontStyle display.
  CODE		This is a new Text node, with a FontStyle which family is set to ["TYPEWRITER", "Courier"].
  COL		This parameter defines visual properties for one or more columns in a 'TABLE' element.
  	align char charoff valign	These parameters define the default parameters for all cells ('TD' elements) for this column and then have to be stored by the converter. They override the values given in the embedding 'COLGROUP' element.
  	span	This defines the number of columns affected by this element parameters.
  	width	This defines the x size of each spanned column and overrides the value given in the embedding 'COLGROUP' element.
  COLGROUP		This element defines a set of columns with default attributes values.
  	align char charoff span valign width	These parameters define the default parameters for all cells ('TD' elements) for this set of columns and then have to be stored by the converter. These attributes can be overridden by the 'COL' element parameters.
  DD		This is a new Text node, with an indentation (contained in a Transform2D with x translation).
  DEL		This is a new Text node, in the continuity of the previous one, if exists. A Form node is needed for align an IndexedLineSet2D onto the text to have a stroke font effect.
  	cite	This implies the node to be included in an Anchor. This is the url of this Anchor.
  	datetime	This is a supplemental information that can be added to the string field of the Text node.
  	Z	This is a format information for the "datatime" field.
  DFN		This is a new Text node, in the continuity of the previous one if exists, with an ITALIC FontStyle display.
  DIR		This is a new Text node, with an indentation (in a Transform2D with x translation).
  DIV		This defines a new Form node.
  	align	This implies that the Form node has constraints with the borders.
  DL		This implies a line feed at opening of this HTML element as well as at the closing, i.e. before the first and after the last MPEG-4 elements converted that defines HTML elements contained in this element.
  DT		This is a new Text node, with a BOLD FontStyle display. The color of this Text can also be changed.
  EM		This is a new Text node, in the continuity of the previous one if exists, with an ITALIC FontStyle display.
  FIELDSET		This cannot be managed by MPEG-4 BIFS. This can be managed by MPEG-J for Keyboard control.
  FONT		This defines a new font style for the following text strings. This should imply then a new Text node, in the continuity of the previous one if exists.
  	color	This defines the emissiveColor of the Material2D of the Shape node in which the Text node in embedded.
  	effect	This implies a new Text node with specific properties. If the value is 'shadow', this new text node is black colored and must be positioned under the current Text node with an interval for shadow effect.
  	face	This is the family field value of the FontStyle.
  	point-size	This is the default size field of the FontStyle. This size is in HTML range (1-7) and then have to be converted in pixel or relative size.
  	size	This increases or decreases the size field of the FontStyle Node. It has then to be recomputed in pixel or relative size.
  	transparency	This defines the transparency of the Material2D of the Shape node in which the Text node in embedded. A factor of 1/100 has to be applied on the HTML value.
  FORM		Basically this element cannot be converted to MPEG-4 at this time because it didn't manage such functionality. In case the KeySensor is normalized, this should be implemented. Another solution is to manage such feature with MPEG-J functionality
  FRAME		This implies a new Layer2D.
  	bordercolor	This defines the emissiveColor of the Material2D of the IndexedLineSet2D nodes defined below.
  	frameborder	If equals to 1, there are IndexedLineSet2D between all Layer2D of the parent FRAMESET.
  	longdesc	This can be ignored. This defines an Anchor node with a info button as geometry on the FRAME's Layer2D.
  	marginheight	This implies a Form node on top of this Layer2D with constraints with top and bottom borders.
  	marginwidth	This implies a Form node on top of this Layer2D with constraints with left and right borders. The Form node can be the same than the one described upper with the added constraints.
  	name	As discussed in the document, when the 'name' or 'target' features are used, this implies all Frame2D to contain an Inline node which DEF name is given by this parameter. This allows changing the URL when the content is updated by a hyperlink.
  	noresize	By default, IndexedLineSet2D defined upper are in the same grouping node than a PlaneSensor2D with constraints on x or y moves. This element implies this PlaneSensor2D not to be added.
  	scrolling	If "YES", geometric object defined for scrollers have to be shown, not if "NO". if "AUTO", their visibility depends on the Layer2D content size.
  	src	This is the address of the content of the top grouping node of the Layer2D. We have to retrieve the HTML content and we can either integrate the converted content directly into the previous top node or add an Inline node to the converted content (recommended).
  FRAMESET		This defines several Layer2D contained in the parent Layer2D node.
  	border	This implies the parent Layer2D to contain IndexedLineSet2D nodes to separate its children Layer2D. Moreover, these IndexedLineSet are at the same level than a PlaneSensor2D vertical or horizontal constraints, depending on the frames position.
  	bordercolor	The IndexedLineSet defined upper have the given color.
  	cols	This defines the size (x-based) fields of the embedded Layers2D.
  	frameborder	If "YES", the lndexedLineSet2D defined upper are to be shown, otherwise not.
  	framespacing	All embedded Layer2D have constraints of position between each other of the given pixel size. This implies resizing of the Layer2D and a Transform2D as parent node.
  	onfocus	The parent Layer2D contains a TouchSensor with ROUTE to its isOver.
  	onblur	The parent Layer2D contains a TouchSensor with ROUTE to its isOver.
  	onload	The parent Layer2D contains a VisibilitySensor with ROUTE to its enterTime.
  	onunload	The parent Layer2D contains a VisibilitySensor with ROUTE to its exitTime.
  	rows	This defines the size (y-based) fields of the embedded Layers2D.
  H1, H2, H3, H4, H5, H6		This is a new Text node, with a fixed size. A carriage return has to be added at the opening as well as at the closing.
  	align	This implies that a Form node, which has constraints with the vertical borders, is the parent of this Text node.
  	align	This implies that a Form node, which has constraints with the vertical borders, is the parent of this Text node.
  HEAD		This delimitates the HTML header from the body (visual part). No visual information is created but initialization may be included.
  HR		This defines a Rectangle which parent Shape node has a LineProperties which color field is '0 0 0' (black).
  	align	This implies that a Form node, which has constraints with the vertical borders, is the parent of this Rectangle node.
  	color	This is the emissivecolor field of the Material2D defined for the Shape node. The default color is '1 1 1' (white).
  	invertborder	This implies the Rectangle to be drawn differently.
  	noshade	This implies the Rectangle to be filled and the remove of the LineProperties.
  	size	This is the y size of the Rectangle.
  	width	This is the x size of the Rectangle.
  HTML		This is the content delimiter. The convertor ignores everything outside this part.
  I		This is a new Text node, in the continuity of the previous one if exists, with an ITALIC FontStyle display.
  IFRAME		This is a Layer2D node.
  	frameborder	If equals to 1, there is a Rectangle with default size as the top of the Layer2D. it is not filled and has a LineProperties.
  	height	This is the y size of the Layer2D.
  	longdesc	This defines an Anchor node with a info button as geometry in the Layer2D.
  	marginheight	This implies a Form node on top of this Layer2D with constraints with top and bottom borders.
  	marginwidth	This implies a Form node on top of this Layer2D with constraints with left and right borders. The Form node can be the same than the one described upper with the added constraints.
  	name	As discussed in the document, when the 'name' or 'target' features are used, this implies all Frame2D to contain an Inline node which DEF name is given by this parameter. This allows changing the URL when the content is updated by a hyperlink.
  	scrolling	If "YES", geometric object defined for scrollers have to be shown, not if "NO". if "AUTO", their visibility depends on the Layer2D content size.
  	src	This is the address of the content of the top-grouping node of the Layer2D. We have to retrieve the HTML content and we can either integrate the converted content directly into the previous top node or add an Inline node to the converted content (recommended).
  	width	This is the x size of the Layer2D.
  IMG		This is a Shape with a Rectangle geometry on which there is an ImageTexture, MovieTexture, CompositeTexture2D (or Layer2D) or CompositeTexture3D (or Layer3D), depending on the URL type. By default, it is an ImageTexture. In case the convertor does not understand the URL content type, it can only display the alternative text in a new Text node.
  	align	This implies that a Form node, which has constraints with the vertical borders, is the parent of this Shape node.
  	alt	This is the alternative Text in case the converter cannot manage the URL. This can also be used at application level for display content information as a "tooltip", as done in most cases for Anchor nodes description.
  	border	In case the value is not 0, this defines that the Shape has a LineProperties with the given pixel size (in fact, the double of it).
  	controls	Can be ignored. In case the URL defines a video clip, we can add some geometrical buttons routed and managed by a MediaControl node.
  	dynsrc	This means the URL defines a movie clip or another VRML (MPEG-4) scene.
  	height	This is the y size of the Rectangle (or Layer).
  	hspace / vspace	This means the ImageTexture is applied to a new Rectangle over this one (and centered) which x and/or y size is decreased of the double of the given space.
  	ismap	The Shape is in the same grouping node than a TouchSensor which touchTime and hitPoint_changed are routed.
  	longdesc	This is the description of the Shape. As for "alt" field, this can be used at application level for display content information as a "tooltip", as done in most cases for Anchor nodes description.
  	loop	In case of a MovieTexture, this defines its loop field.
  	loopdelay	This can be ignored. This implies a TimeSensor and a MediaControl node in case of a MovieTexture.
  	lowsrc	This implies that the URL content has to keep the same ObjectDescriptor that the Shape's one. This content has to be streamed before the main one.
  	onabort	This implies MPEG-J management.
  	onerror	This implies MPEG-J management.
  	onload	The Shape is at the same level than a VisibilitySensor with ROUTE to its enterTime.
  	reload	This indicates (at server side) to the server the frequency at which it has to send the image. This can also be managed on the client side by a MediaControl node combined with a TimeSensor.
  	src	This is the url of the texture of the Shape. The converter has to check whether it can manage the content type or not.
  	start	In case the value contains "mouseover" and the content is a video clip, the Shape is in the same grouping node than a TouchSensor which isOver is routed.
  	transparency	This defines the transparency of the Material2D of the Shape node. A factor of 1/100 has to be applied on the HTML value.
  	usemap	This information is used internally by the converter for keeping consistency on the additional TouchSensors to be applied to the Shape (Cf. MAP).
  	width	This is the x size of the Rectangle (or Layer).
  INPUT		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J for Keyboard control if the converter has such functionality.
  INS		This is a new Text node, in the continuity of the previous one, if it exists. A Form node is needed to align an IndexedLineSet2D onto the text to have an underlined font effect.
  	cite	This implies the node to be included in an Anchor. This is the url of this Anchor.
  	datetime	This is supplemental information that can be added to the string field of the Text node.
  	Z	This is format information for the "datatime" field.
  ISINDEX		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J if the converter has such functionality.
  KBD		This is a new Text node, with a FontStyle which family is set to ["TYPEWRITER", "Courier"].
  LABEL		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J for Keyboard control if the converter has such functionality.
  LEGEND		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J if the converter has such functionality.
  LI		This is a geometrical or textual object in front of a Text node.
  	type	If value is "A", "a", "I", "i" or "1", this is a Text node which string is depending on the order of the element in the list. For "circle", "square" and "disc", this can be a Shape which geometry is respectively a non-filled Circle with LineProperties, filled Rectangle and filled Circle node. A simpler solution is to use specific font, like 'Symbol', which contains already defined characters for such geometric objects.
  	value	This overrides the automatic value increment by setting the string field.
  LINK		This has to be taken into account by the converter for visual aspects default settings (stylesheet, etc ...)
  	charset	Character encoding used for the address. NB: Only UTF-8 is understood by MPEG-4.
  	href	This is the object to load. This can be stylesheet, table of contents, etc ...
  	hreflang	Useful to know if linked content language can be understood by the converter.
  	media	If value is different of "screen" (the default), the whole LINK is ignored by the converter.
  	rel	This indicates the object type. This involves different behavior of the converter. Here are some types: "Alternate": can be used for language switching at application level. "Stylesheet": to be loaded for default size and font settings. Other types can be implemented in the converter.
  	rev	Equivalent to "rel" but for reverse relationship.
  	src	Depending on the relation, this is the URL of the content linked. This can be managed by a button with the relation name in an Anchor node.
  	target	As discussed in the document, when the 'name' or 'target' features are used, this implies all Frame2D to contain an Inline node which DEF name is given by this parameter. This allows changing the URL when the content is updated by a hyperlink.
  	type	Used to specify when "rel" has value of "stylesheet". "text/css" is only relevant here for the convertor. Other values imply additional conversion capabilities (JavaScript, etc ...)
  MAP		This defines an image map. There is no relevant information here. The converter should only keep reference on the "name" for later node creation (Cf. 'AREA'HTML element)
  	name	Has to be stored by the converter for later use.
  MENU		This only delimitates a menu list. The converter should only keep information on display options: add an indentation for all sub- Text nodes.
  	compact	The spacing field of all sub- Text nodes should be decreased.
  META		This is either a Text node (by default) or a WorldInfo node or implies that the parent Layer2D is an Inline node combined with a TimeSensor and Conditional or just some converter status information depending on the "http-equiv" HTML field.
  	content	In case of a Text node, this is to be added to the string field. Otherwise this may contain converter setting information.
  	http-equiv	If "Content-Type", this represents only converter additional information (MIME type). If "Expires", this represents only converter additional information. If "Refresh", this implies that the parent Layer2D is an Inline node combined with a TimeSensor and Conditionnal. The sensor has a cyclelnterval dependending on the content HTML field. The Conditional and Inline URL fields depend also on it. If "Set-Cookie", this can be ignored. This implies application or MPEG-J level management. If this HTML field did not exist, this represents a WorldInfo node.
  	name	In case of a Text node, this is to be added to the string field. Otherwise this may contain converter setting information or WorldInfo additional info field.
  	scheme	This depends on the "http-equiv" and "name" HTML fields.
  NOFRAMES		If the converter has reduced functionalities (with no frame/frameset support), this alternative part has to be taken into account and the given message should be displayed.
  NOSCRIPT		By default this converter did not manage "SGRIPT" tag, then this alternative message has to be displayed to inform the user. An enhanced functionality converter (with JavaScript support -i.e. "SCRIPT" tag management) can ignore this field.
  OBJECT		Depending of the object type, the converter can convert or not the media linked. If it contains non-convertible media, an alternative message is displayed. If the converter has specific conversion capabilities, this element is converted. Such element can contain all kind of media, even media not yet defined that can need extra support or specific plug-in. A specific conversion is then needed.
  OL		This only delimitates an object list. The converter should only keep information on display options: add an indentation for all sub- Text nodes.
  	compact	The spacing field of all sub-Text nodes should be decreased.
  	start	Changes the default starting value (1). Used internally by the converter.
  	type	If value is "A", "a", "I", "i" or "1", all sub- Text nodes string fields are preponned depending on the order of the element in the list. For "circle", "square" and "disc", all sub- Text nodes are preponned with a Shape which geometry is respectively a non-filled Circle with LineProperties, filled Rectangle and filled Circle node. A simpler solution is to use specific font, like 'Symbol', which contains already defined characters for such geometric objects.
  OPTGROUP		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J if the converter has such functionality.
  OPTION		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J if the converter has such functionality.
  P		This is only converter information for a paragraph.
  	align	All sub nodes are included in a Form node, which has constraints with the Layer2D vertical borders.
  PARAM		The conversion of this element depends on the support of 'APPLET' and 'OBJECT' HTML elements by the converter. Default behavior is to ignore this element.
  PRE		This is a new Text node, with no constraint of size or position.
  	cols / width	This has to be taken into account by the converter to cut the string field if needed for the given number of characters.
  	wrap	This implies the converter has to take into account each sentence length to apply wrapping.
  Q		This is a new Text node, or the continuity of the string field of the previous one.
  	cite	This implies to be a new Text node embedded in an Anchor node with the given URL.
  S		This is a new Text node, in the continuity of the previous one, if exists. There is also an IndexedLineSet2D which length is a constraint with the Text length and vertical position with the height. This is done thanks to a Form node.
  SAMP		This is a new Text node, with a FontStyle which family is set to ["TYPEWRITER", "Courier"].
  SCRIPT		A converter that does not manage JavaScript ignores this. This can be converted in an enhanced converter implementation or the converter has to display the NOSCRIPT element alternative message.
  SELECT		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J if the converter has such functionality.
  SMALL		This is a new Text node, with a FontStyle which size is smaller than the normal document one.
  SPAN		This is a new Text node, with display options, which depend on the style that is described.
  STRIKE		Cf. the "S" element.
  STRONG		This is a new Text node, with a FontStyle which style has an additional BOLD style.
  STYLE		This has to be taken into account by the converter for display options. This can be implemented as PROTO and/or DEFed FontStyles or Material2D.
  	media	If value is different of "screen" (the default), we should ignore the whole STYLE.
  	type	If "text/css", this is stylesheet definitions. If "text/javascript", it is ignored by non-JavaScript compliant converters.
  	title	This has to be used has DEF name.
  SUB		This is a new Text node, which FontStyle has a smaller size than the default one. This Text node in embedded in a Transform2D with an increased y translation of half the default FontStyle size.
  SUP		This is a new Text node, which FontStyle has a smaller size than the default one. This Text node in embedded in a Transform2D with an decreased y translation of half the default FontStyle size.
  TABLE		Such an element can be converted by multiple mean. A way to simplify this is to create a Layer2D node embedded in a Transform2D for each TABLE element. Another way to convert this element is to create a Rectangle Shape as top children of a Transform2D node.
  	align	This implies the Layer2D (or Rectangle ) node to be moved in order to be embedded in a new Form node, which has horizontal constraints, defined by this element.
  	background	This defines either the background field of the Layer2D node or the ImageTexture applied to the Material2D of the Rectangle Shape. This defines url field for an image.
  	bgcolor	This defines either the background field of the Layer2D node or the emissiveColor applied to the Material2D of the Rectangle Shape. This defines color for the background.
  	border	In case of use of a Layer2D, this defines a new Rectangle Shape with default size (-2, -2) embedded on top of the Layer2D. This Rectangle's Material2D is not filled and has a LineProperties which width is defined by this element. In case of use of a Rectangle, this one has the same properties than the embedded Rectangle node described before.
  	bordercolor	This defines the lineColor field of the LineProperties node described before.
  	bordercolorlight bordercolordark	This needs specific treament. This implies to replace the LineProperties described before by the use of IndexedLineSet2D Shape nodes. These nodes are embedded in a Form node in order to place them at the extremities of the Layer2D (or the Rectangle ). There are four IndexedLineSet2D (two horizontal and two vertical) on which Form constraints apply (respectively vertically and horizontally). This defines the emissiveColor of the Material2D of these Shape nodes.
  	cellborder	This has to be stored by the converter as pixel width for the boder of all converted cells in the table.
  	cellpadding	This defines constaints for each cell (Rectangle or Layer2D) in the table. The converter has to store this boundary size for later use.
  	cellspacing	This defines border size for each cell (Rectangle or Layer2D) in the table. The converter has to store this border size for later use.
  	cols	This parameter implies the converter to compute each column width depending on the table width and on the number of columns.
  	frame	This parameter implies visual changes about which parts of the table will be visible. The converter has to take into account this information. This may need specific treatment for border visibility. Like for the 'bordercolorlight/dark' parameters, the use of IndexedLineSet2D may be necessary.
  	gradangle gradcolor	This implies the use of an IndexedFaceSet2D in the replacement of the Layer2D's top Rectangle (in case of the use of a Layer2D ) or in the replacement of the Rectangle defining the table. This IndexedFaceSet2D is defining the table (rectangle shape) layout. These parameters define the color field of this node and the converter has to compute the colors for each Coordinate2D (coord field) depending on these values.
  	height width	This defines the y/x size of the Layer2D or Rectangle defining this table.
  	href	This implies all converted table content to be embedded in an Anchor node, which url is defined by this parameter.
  	hspace vspace	This defines a new Form node embedding the Layer2D or Rectangle defining the table, which has horizontal/vertical constraints with media surrounding the table.
  	nowrap	This parameter implies the converter not to recompute the x size of the table in case it extends beyond the embedding Layer2D visibility.
  	rules	This implies visual changes for the table. The converter has to store this parameter in order to use it later for cell border visibility.
  	summary	This can be ignored because the converter converts visual information. In case of conversion, the converter can display a new Text node under or over the converted table.
  	transparency	Like for the border parameter, this implies the use of an embedded Rectangle in case of the use of a Layer2D as table conversion. There is no additional node in case of use of a Rectangle as table conversion. This defines the Material2D transparency field of this Rectangle Shape. A division of 100 of the HTML value is necessary for range compatibility.
  TBODY		This defines a body part of the table. This implies the converter to check table parts order (header, body and then footer) in order to visually order them.
  	align	This implies the converter to compute width of cells and the use of a new Form node that containts all cells in this part. These Form node constraints are defined by this parameter.
  	char charoff	These define visual parameters. The converter can have to insert a new Transform2D for each cell in order to visually center them relating to a specific character with an offset defined by 'charoff'.
  	valign	This defines FontStyle justify field default value for all Text nodes embedded in the table.
  TD		This defines a cell in a table. This can be converted by a Layer2D but should be better converted by a Rectangle node.
  	abbr	This defines a abbreviated version of the cell content. The converter has to compute cell size in order to check whether it needs to display this information or the complete cell content.
  	absheight abswidth	These parameters define the absolute y/x sizes of the Rectangle or Layer2D defining the cell. If the computed cell content size doesn't fit in the cell, part of it can be invisible. As these parameters are not used anymore, the converter can ignore them.
  	align	This implies the converter to compute width of cells and the use of a new Form node that embeds all cell contents. These Form node constraints are defined by this parameter.
  	axis	This defines non-visual data and is moreover not fully normalized in HTML. The converter should then ignore this parameter.
  	background	Idem than 'background' parameter of 'TABLE' element applied to the cell.
  	bordercolorlight bordercolordark	Idem than the same parameters for 'TABLE' element applied to the cell.
  	char charoff	Idem than the same parameters for 'TABLE' element applied to the cell.
  	colspan rowspan	These indicate that this cell can span from 1 to all columns/row of the table. The converter must store this information in order to display the cell content as required, i.e. spanned to the given number of columns/rows.
  	gradangle gradcolor	Idem than the same parameters for 'TABLE' element applied to the cell.
  	height width	These parameters define the absolute y/x sizes of the Rectangle or Layer2D defining the cell.
  	headers	This defines non-visual data and is moreover not fully normalized in HTML. The converter should then ignore this parameter.
  	maxlines	This parameter defines the maximum number of lines for this cell. If the computed cell content lines number do not fit in the cell, part of it can be invisible. As this parameter is not used anymore, the converter can ignore it.
  	nowrap	This is visual information the converter has to take into account. This defines that the cell size must match its content size, whatever big it is.
  	scope	Cf. 'headers' attribute.
  	transparency	Idem than the same parameter for 'TABLE' element applied to the cell.
  	valign	Idem than the same parameter for 'TBODY' element applied to the cell.
  TEXTAREA		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J if the converter has such functionality.
  TFOOT		Cf. 'TBODY' HTML element applied to footer information.
  TH		Cf. 'TD' HTML element applied to header information. The converter has to display these information before 'TD' elements. A "BOLD" style value of FontStyle for embedded Text nodes should be used.
  THEAD		Cf. 'TBODY' HTML element applied to header information.
  TITLE		This information has to be used as the title field of the WorldInfo node.
  TR		This defines a row in a table and contains the 'TH' and 'TD' HTML elements. The converter has to store this row definition for visual parameters.
  	align background bordercolorlight bordercolordark char charoff nowrap transparency valign	These parameters define the default parameters for all cells ('TD' elements) for this row and then have to be stored by the converter.
  	bgcolor	This redefines the 'bgcolor' of the embedding 'TABLE' element.
  TT		This is a new Text node, with a FontStyle which family is set to ["TYPEWRITER", "Courier"].
  U		This is a new Text node, in the continuity of the previous one, if it exists. There is also an IndexedLineSet2D which length is a constraint with the Text length and positioned under it. This is done thanks to a Form node.
  UL		This only delimitates an unsorted list. The converter should only keep information on display options: add an indentation for all sub-Text nodes and precedates the first one with a carriage return.
  VAR		This is a new Text node, with a FontStyle which style is set to "ITALIC".

Annex 2

• XHTML/HTML proprietary elements	MPEG-4 node - method to convert
  AUDIOSCOPE		This is platform dependant and then should be ignored.
  BGSOUND		In case the linked media is understandable by an MPEG-4 platform, this is a Sound2D node with a unique AudioClip.
  	loop	This is the loop field of the AudioClip node.
  	src	This is the url field of the AudioClip node.
  BLACKFACE		This is a new Text node, with a FontStyle which style is set to "BOLD". Its parent Shape node contains a LineProperties.
  BLINK		This is a new Text node, in the continuity of the previous one if it exists. There is a need of a default looping TimeSensor and a Switch node to switch on and off the Text rendering.
  BQ		Cf. standard BLOCKQUOTE HTML element.
  COMMENT		This defines an HTML comment. Just ignore it or add the comment as MPEG-4 commentary syntax.
  EMBED		The node creation depends on the embedded object content type. This can be audio, Java applet, video, VRML, etc ... Then some content can be easily converted, some not at all, because requiring proprietary plug in. In the case the converter can manage it, it creates a Layer2D or Layer3D.
  	align	The Layer(2/3D) is included in a Form node which has constraints with the parent Layer2D vertical borders.
  	border	In case the value is not 0, this defines that there is a top Rectangle Shape that has a LineProperties with the given pixel size (in fact, the double of it).
  	frameborder	If equals to 1, there is a Rectangle with default size as the top of the Layer. it is not filled and have a LineProperties.
  	height	This is the y size of the Layer2/3D.
  	hidden	If TRUE this layer shouldn't be created.
  	hspace	This is a value for alignment in pixel for horizontal constraints with the borders in the Form node at the top of the Layer2/3D node and for all nodes in it.
  	name	This has to be used as DEF name for the Layer content top grouping node.
  	palette	Just ignore this field.
  	pluginspage	This should be displayed as a Text node in an Anchor node to help user in case the content is not convertible.
  	pluginurl	Just ignore this field.
  	src	This implies an Inline node at the top of the Layer with the given URL.
  	type	This helps the converter to know if the content type is convertible or not.
  	units	Has to be used by the converter for size conversions.
  	vspace	This is a value for alignment in pixel for vertical constraints with the borders in the Form node at the top of the Layer2/3D node and for all nodes in it.
  	width	This is the x size of the Layer2/3D.
  FN		This is a Text node, in the continuity of the previous one if exists. This can also be ignored because this is a platform specific element.
  	href	This implies the Text node to be included in an Anchor node.
  	id	This has to be used as DEF name for the text.
  ILAYER		This defines a new Layer2D.
  	above	This is to be used by the converter to order the ILAYERs one above each other.
  	background	This is the image to put in the Background2D node (url field) of this layer.
  	below	This is to be used by the converter to order the ILAYERs one below each other.
  	bgcolor	This defines the backColor to be used in the Background2D node of this layer.
  	clip	This defines the size of the Layer2D. This also implies a Transform2D as parent node of this layer.
  	height	This defines the y size of the Layer2D.
  	id	As discussed in the document, when the 'name' or 'target' features are used, this implies all Frame2D to contain an Inline node which DEF name is given by this parameter. This allows to change the URL when the content is updated by an hyperlink.
  	left	This implies a Transform2D as parent node of this layer. This field defines the x translation field.
  	name	Cf. "id" HTML field.
  	onfocus	The parent Transform2D contains a TouchSensor with ROUTE to its isOver.
  	onblur	The parent Transform2D contains a TouchSensor with ROUTE to its isOver.
  	onload	This Layer2D contains a VisibilitySensor with ROUTE to its enterTime.
  	onmouseout	The parent Transform2D contains a TouchSensor with ROUTE to its isOver.
  	onmouseover	The parent Transform2D contains a TouchSensor with ROUTE to its isOver.
  	pagex	This implies a Transform2D as parent node for this layer. This field defines the x translation field. Moreover, the Transform2D is a top node of the top-grouping node (Layer2D).
  	pagey	This implies a Transform2D as parent node for this layer. This field defines the y translation field. Moreover, the Transform2D is a top node of the top-grouping node (Layer2D).
  	src	This Layer2D contains an Inline node, which link on the given URL.
  	top	This implies a Transform2D as parent node for this layer. This field defines the y translation field.
  	visibility	This is information for the converter to decide whether the Layer2D has to be drawn or not.
  	width	This defines the x size of the Layer2D.
  	z-index	This is to be used by the converter to order the ILAYERs.
  KEYGEN		MPEG-4 BIFS cannot manage this. This can be managed by MPEG-J if the converter has such functionality.
  LAYER		Identical to ILAYER. The only difference is the default positioning.
  LIMITTEXT		This has to be used internally to limit the number of character of each sentence.
  	size	This is the maximum number of displayable characters.
  	value	This is the Text node to which the restriction has to be applied.
  	width	This has to be used as length for all Text nodes.
  LISTING		Cf. standard PRE HTML element.
  MARQUEE		This defines a new Text node in a parent Transform2D on which a Positionlnterpolator2D is applied. There is also a Rectangle at the same level than the Transform2D. An additional TimeSensor is necessary for the animation process.
  	align	This defines the y translation of the Transform2D.
  	behaviour	This defines the values inside the Positionlnterpolator2D. For example, the 'ALTERNATE' value defines a bounce movement of the text from left to right and vice-versa.
  	bgcolor	This defines the emissiveColor field for the Shape of the Rectangle.
  	direction	This defines the values inside the Positionlnterpolator2D.
  	height	This defines the y size of the Rectangle and default y translation of the Transform2D.
  	hspace	This implies a Form node over the Transform2D with constraints on the vertical borders.
  	loop	This defines the loop field of the TimeSensor directing the Positionlnterpolator2D.
  	scrollamount	This defines the values inside the Positionlnterpolator2D.
  	scrolldelay	Can ignore this. This is renderer dependant.
  	transparency	This defines the transparency field of the Shape embedding the Rectangle.
  	vspace	This implies a Form node over the Transform2D with constraints on the horizontal borders.
  	width	This define the x size of the Rectangle.
  MULTICOL		This defines either a Form node or a Layer2D node. This is an implementation choice.
  	cols	This defines the number of vertical layers of a horizontal constraint for the Form.
  	gutter	This defines either y translation for Transform2D over the layers or an additional horizontal constraint for the Form.
  	width	This defines either y size for the layers or an additional horizontal constraint for the Form.
  NOBR		This is only visual setting for the converter in order no to break up lines which length is bigger than the display width. This needs specific string length computation in order to know whether the converter has to break up the sentence within lines.
  NOEMBED		This is the alternative Text node that the converter has to display when the embedded content type is not convertible.
  NOSMARTQUOTES		This is platform dependant. The converter can ignore this.
  PLAINTEXT		This is a converter setting in order not to interpret embedded HTML elements. Instead of interpreting them, they have to be included in the string field of the current Text node.
  SERVER		This should be ignored. It can be however implemented in a specific manner on the server side.
  SHADOW		This defines a new Text node placed "under" the current Text node which emissiveColor is black in order to create a shadow effect.
  SIDEBAR		This defines a new Layer2D. This feature is no longer supported so the converter should ignore this element and display the media within it as it.
  	align	This defines the x translation of the layer's parent Transform2D.
  	width	This defines the x size of the Layer2D.
  SPACER		This can define either a Form node or a new Transform2D node, which embeds the HTML element converted containt.
  	align	This defines constraints for the constraints field of the Form node or the translation field of the Transform2D.
  	height	This define vertical constraints for the Form node or y translation of Transform2D.
  	type	If equals to "VERTICAL" or "HORIZONTAL", this defines respectively vertical and horizontal constraints for the Form node or y and x translation of Transform2D. If equals to "BLOCK", there is a Rectangle as children of the Form which size depends on the "width" and "height" element fields.
  	width	This defines horizontal constraints for the Form node or x translation of Transform2D.
  WBR		This is a visual setting for the converter, which makes it break a line at the given position if the display width is not sufficient to contain the whole sentence.
  XMP		Idem as the PLAINTEXT element.

Annex 3

• CSS elements	MPEG-4 node - method to convert
  azimut		This defines default fields values for Sound and Sound2D nodes, as well as other audio nodes.
  background		This defines the default layers' background field. If "transparent" is set, there is no Background2D nor Background node.
  	background-color	This defines the skyColor of Background nodes and backColor of Background2D ones.
  	background-image	This defines the URL fields of Background nodes and url field of Background2D nodes.
  	background-repeat	If this is different from "no-repeat", Background or Background2D nodes are replaced by a top Rectangle node with default size (-2, -2).
  	background-attachment	If "fixed", when scrollers are displayed, they don't have ROUTES to the upper Layer2D parent Transform2D.
  	background-position	If different from "top left" (default), Background or Background2D nodes are replaced by a top Bitmap or Rectangle node.
  border border-left border-right border-bottom border-top		This defines default parameters for all object borders i.e. IndexedLineSet2D and LineProperties.
  	border-... -width	This defines the width field of LineProperties.
  	border-...-style	This defines the lineStyle field of LineProperties nodes.
  	border-...-color	This defines the lineColor field of LineProperties nodes.
  	border-spacing	This defines the translation field for Transform2D nodes between two adjacent borders.
  	border-collapse	This implies a visual aspect for cells display, and then the translation field for Transform2D nodes between two adjacent cells.
  caption-side		This implies a visual aspect for the table captions, and then the translation field for Transform2D nodes between the caption and the first cells.
  clear		If different from 'none', the converter has to check whether there are elements on sides of the given object. If this is the case, it has to translate the current object in order the constraint to be applied.
  clip		This cannot be managed by the current version of MPEG-4.
  color		This defines the default color of an XHTML/HTML text element. These values are to be stored by the converter for later use.
  content		This defines an Inline node, which address is given with the url element, or specific content, which is described in the string element, to be put before of after the specified element.
  counter-increment		This implies the converter to internally store counters for lists or identifiers. This increases the counter of 1.
  counter-reset		This implies the converter to internally store counters for lists or identifiers. This resets the counter (equals it to 0)
  cue		This defines new Sound2D and AudioClip nodes.
  	cue-before cue-after	This implies a TimeSensor node between the two objects.
  cursor		This cannot be managed by the current version of MPEG-4. This should be managed at application level.
  direction		This is the default value of leftToRight field for FontStyle nodes.
  display		This defines what the HTML element should display.
  elevation		These are spatial parameters for audio nodes.
  empty-cells		This is a visual property the converter has to store. This defines the visibility of empty cells.
  float		This defines constraints for the positioning of object with respect to the other thanks to Form nodes. Some of these constraints can be impossible to convert.
  font		This defines default FontStyle values.
  	font-style	This defines the style field of the FontStyle node.
  	font-variant	In case the value is 'small-caps', this implies decrease of the size field of the FontStyle node for uppercases characters.
  	font-weight	This defines the style field of the FontStyle node. Some values are impossible to convert.
  	font-size	This defines the size field of the FontStyle node.
  	line-height	This defines the spacing field of the FontStyle node.
  	font-family	This defines the family field of the FontStyle node.
  @font-face		This defines a new font face. Basically, it is impossible to create new font faces in MPEG-4. Then, only existing fonts should be taken into account. The default font face should replace others.
  font-size-adjust		This is not manageable in MPEG-4.
  font-stretch		This implies the Text node to be embedded in a Transform2D with a x scale factor. This can also be implemented in changing the length field of the Text nodes.
  height		This defines the default height for the current element. Basically, MPEG-4 cannot manage this constraint. This can be implemented as a new Layer2D which size is the given value.
  @import		This implies the converter to download the given URL in order to retrieve another stylesheet definition document.
  left right bottom top		These define constraints of Form nodes for geometrical nodes relatively to their parent containing node. In case the value is different from 'auto', this then implies that all geometrical nodes are embedded in a Form node.
  letter-spacing		This is basically not a feature of MPEG-4. This can be implemented in the way the converter manages the text character by character in Text nodes.
  line-height		This is the minimum spacing field value for FontStyle.
  list-style		This defines the list style for listing HTML elements.
  	list-style-type	This defines the type of the list: either a geometrical object ( Shape node) or a textual ( Text node).
  	list-style-position	This defines new Transform2D and Form nodes for all embedded text lines.
  	list-style-image	This defines an ImageTexture as list item marker.
  margin margin-top margin-bottom margin-left margin-right		This defines the width for all margins (LineProperties and IndexedLineSet2D).
  marker-offset		This implies a new Form node with constraints between list markers and following Text.
  max-height min-height max-width min-width		These are vertical and horizontal constraints for all geometrical nodes, which have to be embedded in a Form node.
  orphans widows		This defines the number of lines the converter has to put at the end and/or the begin of the Layer2D.
  outline		These define the default values for LineProperties.
  	outline-color	This defines the lineColor field.
  	outline-style	This defines the lineStyle field. Some styles can be non-convertible.
  	outline-width	This defines the width field.
  overflow		This cannot be basically managed by MPEG-4. A partial implementation can be to use additional Layer2D to clip the content.
  padding padding-top padding-bottom padding-left padding-right		This defines constraints for new Form nodes for geometrical nodes.
  page		This defines a new Layer2D and IndexedLineSet2D.
  page-break before page-break-after page-break-inside		This defines new IndexedLineSet2D. with Form constraints.
  pause pause-after pause-before		The converter should ignore this. This can be implemented with BIFS update commands.
  pitch		This defines default pitch field value for AudioClip nodes.
  play-during		This defines the loop field for AudioClip nodes.
  position		This defines constraints for geometrical nodes embedded in Form nodes.
  quotes		This is an information the converter has to store as default HTML 'Q' element quotes.
  size		This is an information the converter has to store for later page break purpose.
  table-layout		This defines the way the table cells are drawn. This is converter visual information.
  text-align		This defines new Form nodes with horizontal constraints with the vertical borders.
  text-decoration		This defines the way text is displayed. Possible values are: 'none': nothing more is required than the Text node. 'underline', 'overline', 'line-through': an IndexedLineSet2D and Form nodes will be needed to put a line respectively under, over and on the middle of the Text node (vertically).
  text-indent		This can be implemented either by a parent Transform2D node with a x translation or by a parent Form node with constraints with the Layer2D left border.
  text-shadow		This defines a new Text node, which string field is the same. This new Text node has a parent Transform2D with a translation of the given position. The emissiveColor of the Shape can be also provided.
  text-transform		This is an information the converter has to store in order to transform some Text string characters into uppercase or lowercase.
  unicode-bidi		This defines the text format. MPEG-4 only manages UTF8 characters.
  vertical-align		This defines constraints for Form nodes embedding geometrical objects. Some constraints may be non-convertible.
  visibility		This is an information the converter has to store in order to display or not the content.
  white-space		This is an information the converter has to store as text behavior. This can imply the converter to manage Text nodes character by character.
  width		This defines constraints for geometrical objects embedded in a Form node.
  word-spacing		This is basically not a feature of MPEG-4. This can be implemented in the way the converter manages the text word by word in Text nodes.
  z-index		This implies either a new OrdererGroup node or the converter has to re-order the children nodes to follow the drawing index order.

Claims (9)

1. Device to treat HTML information including data and at least one page and its associated frame layout, comprising

   means to generate a structure of the HTML data,

   characterized in that it comprises:

   means to convert the HTML page and its associated frame layout into MPEG4 layout,

   means to convert the HTML data into MPEG-4 nodes.
2. Device according to claim 1 characterized in that the means to generate a structure of the HTML data are able to generate a tree structure of the HTML data and that the conversion means intended to convert the HTML data into MPEG-4 nodes convert the HTML tags and the cascading style sheets to Binary Information For Scene information.
3. Device according to any of claims 1 or 2 characterized in that the HTML data are standardized HTML data or browser-specific HTML data.
4. Device according to any of claims 1 or 3 characterized in that it comprises means to generate an output document (19) from the obtained MPEG-4 layout and MPEG-4 nodes, said output, document being described using the MP4 format or the XMT format.
5. Multimedia system characterized in that it comprises:

   at least one browser (3) to generate a request of at least one HTML document (9) located on a remote server (5),

   at least one device (1) according to any of claims 1 to 4,

   at least one transmission link (6) between said remote server (5) and said device (1) according to claim 1 for transferring said requested HTML document (9),

   at least one display (4) to display said MPEG-4 document (19),

      said device according to claim 1 to convert said requested HTML document (9) into an MPEG-4 document (19) before transferring it to said display (4).
6. Multimedia system characterized in that it includes:

   a server (7),

   a device for displaying multimedia data (2),

   a device for receiving a request for an HTML document from a user (3),

   a transmission link (8) being able to for connecting said server (7) and said device (3) for receiving the request for transmitting said request to said server (7),
      said server (7) including a device according to any of claims 1 to 4 for converting said HTML document (9) into an MPEG-4 document (19), said transmission link (8) transmitting said HTML document (9) converted into MPEG-4 to the device (2) for displaying multimedia data.
7. Multimedia system according to one of claims 4 to 5 characterized in that the server is intended to be connected to the World Wide Web (5) and in that said requested HTML document (9) is located on the World Wide Web (5).
8. Method for treating HTML information including at least one page and its associated frame layout wherein

   one generates a structure of the HTML data,

   characterized in that

   one converts the HTML page and frame layout into MPEG4 layout,

   one converts the HTML data into MPEG-4 nodes.
9. Computer program product comprising portions of software code or program instructions adapted to implement the method of claim 8 for converting HTML data into MPEG-4 data, when the computer program is executed on a computer.

Images